WEBVTT

1
00:00:01.640 --> 00:00:04.640
Artificial intelligence is rapidly
changing the world of work.

2
00:00:04.640 --> 00:00:08.360
Pia Papke and Jan-Ove Becker from
Littler explain how companies

3
00:00:08.360 --> 00:00:11.480
can make legally compliant and
efficient use of the new possibilities.

4
00:00:11.480 --> 00:00:15.480
Not long after ChatGPT was
launched in November 2022,

5
00:00:15.480 --> 00:00:18.800
a technology company approached us
and said

6
00:00:18.800 --> 00:00:22.080
My employees have
suddenly started using ChatGPT

7
00:00:22.080 --> 00:00:24.560
and other tools
that are available.

8
00:00:24.560 --> 00:00:27.360
And they weren’t sure
Is that actually ok?

9
00:00:27.360 --> 00:00:29.760
Can we do that?
Should we regulate it?

10
00:00:29.760 --> 00:00:33.520
And then we very quickly got in touch
with our American colleagues,

11
00:00:33.520 --> 00:00:36.120
who already have a big
practice group in this area,

12
00:00:36.120 --> 00:00:39.480
which we have now rolled out worldwide,
and we asked,

13
00:00:39.480 --> 00:00:43.200
“What guidelines have you already drafted
so that companies can use this legally?”

14
00:00:43.200 --> 00:00:46.160
There are various areas of
regulation that apply,

15
00:00:46.160 --> 00:00:49.040
which were then also very important
for the clients at that time.

16
00:00:49.040 --> 00:00:54.640
Of course, one thing that is totally relevant
is which tools do we want to use in the company

17
00:00:54.640 --> 00:00:58.600
and which ones do we want to use more?

18
00:00:58.600 --> 00:01:01.920
For example, do my
employees use ChatGPT,

19
00:01:01.920 --> 00:01:04.680
or are there other tools
that we might be using?

20
00:01:04.680 --> 00:01:08.840
Then it is very important
to check again afterwards

21
00:01:08.840 --> 00:01:14.960
what content has been generated by the AI,
i.e. whether any discrimination

22
00:01:14.960 --> 00:01:19.120
has occurred, for example.
There are often problems here with AI.

23
00:01:19.120 --> 00:01:23.040
After all, the AI only learns
what it is fed.

24
00:01:23.040 --> 00:01:26.880
And if this content is discriminatory,
then it can also happen that the AI

25
00:01:26.880 --> 00:01:31.680
generates discriminatory results
and then in turn also checks them for accuracy.

26
00:01:31.680 --> 00:01:35.400
So the AI is only as smart
as the level of knowledge it is fed with.

27
00:01:35.400 --> 00:01:36.960
And so one thing is very important:

28
00:01:36.960 --> 00:01:41.280
always check the results and define
checking obligations for employees.

29
00:01:41.280 --> 00:01:44.800
Also, if you look at the statistics
and why companies actually want to start using AI,

30
00:01:44.800 --> 00:01:47.760
these are not generally
negative reasons,

31
00:01:47.760 --> 00:01:50.360
but rather, as you just mentioned,
it is actually about

32
00:01:50.360 --> 00:01:54.360
making the work easier,
simplifying processes and perhaps also

33
00:01:54.360 --> 00:01:57.760
relieving people of things
that are also arduous.

34
00:01:57.760 --> 00:02:01.960
So writing emails, small things,
that’s something a computer can do well for me.

35
00:02:01.960 --> 00:02:05.080
And that has many advantages.
If you put fears in the foreground,

36
00:02:05.080 --> 00:02:07.720
you can also virtually eliminate them.
On the other hand:

37
00:02:07.720 --> 00:02:11.000
Many years ago, MTV opened
with “Video Killed the Radio Star”

38
00:02:11.000 --> 00:02:14.280
and I don’t think we were
all that ungrateful

39
00:02:14.280 --> 00:02:18.240
that the music side has also evolved
and I think the same applies to AI.

40
00:02:18.240 --> 00:02:20.400
The experts at Littler
make it possible for companies

41
00:02:20.400 --> 00:02:22.200
to keep up with
technological developments

42
00:02:22.200 --> 00:02:25.040
and legally implement AI tools
in their operations.